, ,

The Technical Case for Mixing Cloud Computing and Manufacturing

Technical Case for Mixing Cloud Computing and Manufacturing

The Technical Case for Mixing Cloud Computing and Manufacturing

By David S. Linthicum for Nelson Hilliard

Manufacturing, as a vertical, is the market sector with the most potential to be improved by cloud computing.  Today’s manufacturing processes often lack visibility into resource consumption metrics, productivity, and even logistics.  The end result is that huge blind spots in the building process often means a lack of productivity and efficiency that leads manufacturing companies down unprofitable paths. 

“For American manufacturers, the bad years didn’t begin with the banking crisis of 2008.  Indeed, the U.S.  manufacturing sector never emerged from the 2001 recession, which coincided with China’s entry into the World Trade Organization.  Since 2001, the country has lost 42,400 factories, including 36 percent of factories that employ more than 1,000 workers (which declined from 1,479 to 947), and 38 percent of factories that employ between 500 and 999 employees (from 3,198 to 1,972).  An additional 90,000 manufacturing companies are now at risk of going out of business.” [1] 

The answer to manufacturing growth in the US is more about technology and intelligence than it is about economic issues.  This has come to light within several manufacturing companies that leverage cloud computing, IoT, and big data analytics to find a path to a competitive advantage that foreign low-cost competitors will find hard to replicate. 

It’s About Real-Time

At the heart of manufacturing issues is the lack of real-time visibility into the state of the business at any given time.  Manufacturing case studies that I’ve been involved with have very little understanding, as to the state of the business, including production rates, quality issues, distribution, etc..  These enterprises basically become businesses that respond to behaviour, such as missing production runs.  Instead, they need to become businesses with the ability to make proactive adjustments to manufacturing processes, while leveraging predictive analytics. 

The notion of real-time data analytics dates back to the concept of the real-time enterprise (RTE).  It’s also known as the ability to monitor what’s going on in a business, as it happens.  Early on, this notion related more to data integration than to cloud, data, and IoT.  These days it’s about how you externalize the data, and what you do with it as well, in terms of making analytical sense of the information become a part of the manufacturing process. 


In the next section, we’ll discuss the impact of cloud computing in general, as well as how it does not introduce any new concepts, such as real-time or predictive analytics.  Predictive analytics, when combined with the notion of real-time data possessing, and the efficiencies of the public cloud, is where the real value lies. 

In manufacturing companies, traditional data warehousing technology allowed them to slice and dice data that was used to make decisions.  However, by the time the data was abstracted from operational data stores, rolled up and analyzed, weeks or months had passed.  Those who operate the business had no ability to correct issues as they occurred.  The ability to see current data, and respond to current analytics based upon that data, means the difference between having a manufacturing operation that can make margins of 50-60 percent, versus margins of only 10-15 percent, which means the business is dying. 

These days, we have the cloud and Big Data technology to introduce these opportunities.  Emerging cloud databases can be provisioned on-demand to provide access to Big Data technology with the horsepower to provide real-time data analytics.  Real-time data analytics means that the data gathered from operational data stores is almost immediately analyzed and externalized to humans, who can act upon information that is seconds old.

For instance, a plant manager could monitor the current maintenance needs of the plant equipment and thus predict downtime and lost production hours with a great degree of accuracy.  In the healthcare world, more people wearing health telemetry devices will lead to real-time data analytics that could predict, for example, heart attacks based upon the current data being produced.  More people wearing health telemetry devices can also show historical data that can illustrate patterns of data that indicate the likelihood that a negative health event could occur.[2]


This “perfect storm” of cloud, big data, analytics, and IoT has raised the bar, in terms of what is possible to on the path to better manufacturing.  Moreover, it has the potential to be a game changer, in terms of making a once competitive US manufacturing sector once again relevant.  But, it can also make worldwide manufacturing a much more efficient and less expensive process.  This makes goods cheaper, and thus more obtainable by all.  We could be looking at quality cars that cost less than ten thousand dollars, one hundred dollar flat screens, and two hundred dollar major appliances.  This is due to automation and intelligence, rather than cheaper labor or poor working conditions. 


How Real-time Metrics Work in Manufacturing


The notion of real-time metrics is to provide real-time visibility into data points – both raw and calculated on the fly – that provide required operational information.  When this notion is applied to manufacturing, the data points coming from IoT-enabled devices, operational systems, and other points that generate data will provide the raw data needed to determine real-time analytics.  These analytics can be embedded into business processes for automated reactions, or externalized to humans who can take corrective actions. 

Key to the inner workings of real-time metrics is the concept of service aggregation or data aggregation.  These mechanisms allow users to combine many measurements into a single compound or complex measurement (see the following figure).  IT can define what’s exposed to the dashboard, and to those who view the metrics.  The information can be customized so users can see and understand the metrics gathered from the measurements being made in real-time. 

We can do a couple of things with these aggregations: 


First, we can make sure humans are alerted to specific issues so problems can be corrected.  For instance, an alert about a manufacturing machine on the plant floor that does not provide the output needed to supply the other parts of the manufacturing process.  Because production is limited by the output of that device, it needs to be repaired ASAP.


Second, and most important, the aggregations can be embedded into real-time processes, such as production monitoring systems, and automated corrective actions can be taken.  For example, the system can automatically order raw materials that are needed for a production run, based upon predictions that are made using data gathered off of the plant floor.  This means that production will increase with the right materials available that are aligned pretty much with the overall production rate.  Missing runs due to the lack of raw materials are common problems in manufacturing. 


Other examples would be the ability to understand the calculated risk that a device will go into failure (such as a component in a metal press), based upon measurements coming in that are run through a special risk analytics algorithm, which is the composite or calculated value.  Since this occurs on the fly, the data displayed on the dashboard considers the raw data coming in from the device, as well as a pattern that’s processed in real-time to add value to the meaning of the data, such as the likelihood that a failure will occur.  The real-time metrics user can even place those calculations around policies or limits that can perhaps create an alert if the aggregated value goes out of a pre-set range. 

Aggregation of incoming data provides views that make more sense than raw data views

The Technical Case for Mixing Cloud Computing and Manufacturing

Source: Cloud Technology Partners[3]

Enter The Cloud

Cloud is the game changer here, not because cloud computing brings new technologies, as stated above, but because it makes technology affordable to the manufacturing vertical that had traditionally thin margins, and thus economized on IT systems.  The cloud makes access to critical systems affordable, and the elastic nature of cloud computing and “pay by usage” billing means that no high up-front investment needs to be made, and the cloud bill grows only with the growth of the business. 

 The Mint Jutras firm completed a survey of SaaS adoption in manufacturing, distribution and other industries.  They found the following:

 – 49% of respondents in the manufacturing & distribution industries do not understand the difference between single- and multi-tenant SaaS architectures.  Overall, 66% of respondents to the survey were unaware of the difference.

 – SaaS-based applications are 22% of all manufacturing and distribution software installed today and will grow to 45%.

 – The three most important characteristics of a SaaS solution in manufacturing and distribution include giving customers a measure of control over upgrades, consistent support for global operations, and allowing for rapid and frequent upgrades.[4]

 The long and the short of manufacturing and the cloud is that the manufacturing vertical has the most to gain from the use of the public cloud.  That said, as revealed in the results of the survey above, they are likely to have a little understanding of what cloud is, and what it can do.  However, the use of SaaS continues to grow in-line with other industries. 

The movement in manufacturing needs to be around the growth of IaaS usage, including cloud-delivered servers, databases, data integration, and other core components needed to provide the types of services listed earlier in this article.  AWS provides all of these components, as does Google and Microsoft. 

That said, what keeps many manufacturing companies out of the cloud is the lack of skills and knowledge.  It takes a specific skill set to properly integrate existing ‘some-time’ systems that provide no real-time visibility or automated responses with new cloud-based systems that provide the ability to operationalize new and existing data points.  The objective is to provide a quick ROI, as well as the ability to move operations in more productive and less expensive directions. 

The fundamentals are well understood and are becoming easier to understand by the manufacturing organizations.  What’s missing is a stepwise approach that spells out the cloud conversion approach with enough detail to provide the company with a path to tactical and strategic success.  Here’s my recommended approach:

Step 1:  Understand what data is currently available, and the frequency with which it’s externalized.  Many manufacturing organizations rely upon pre-canned applications that produce only a few data points.  While these are not much help when moving to cloud-based real-time systems, they are a starting point. 

Step 2:  Define the data points that can be externalized from the plant floor.  These are the machines, the employees, IoT, RFID tags, etc., that allow you to track items that were previously untracked.  There are more out there than you know since most manufacturing equipment now comes with data integration features, most of which are not used. 

Step 3:  Design the logical architecture, including data collection, data analysis, and data intelligence, all within a cloud-based system.  Do not choose the physical technology, but understand the logical requirements that are needed to define your target cloud-based system. 

Step 4:  Pick your technology.  This includes cloud providers, subsystems you will leverage, and use cases for each.  This is the most detailed and longest part of the process and carries the most risk.  Take your time and make sure to test the technology you’ve selected. 

Step 5: Consider security and governance.  While most manufacturing organizations consider security to be an afterthought, hackers can screw things up to a point of putting you out of business.  Make sure security is systemic, as well as resources, APIs, and data governance. 

Step 6:  Make the improvement process ongoing.  All of the analytics and processes should be under constant improvement (continuous improvement), and that makes changes easy to implement.  You do this by placing system volatility into a domain, thus changes are made using configuration patterns versus redevelopment and deployment. 

The process of moving manufacturing organizations to the cloud is underway right now.  That is, for the companies that have the right vision.  Some will claim that the use of the cloud is too expensive for cash-strapped manufacturing companies.  I would point toward the growing evidence that the companies who make the investment actually save money, so that’s not an argument in my book.

[1] http://prospect.org/article/plight-american-manufacturing

[2] http://www.rtinsights.com/iot-meets-real-time-data-analytics/

[3] https://gigaom.com/report/real-time-metrics-for-improved-business-operations-and-agility/

[4] http://www.forbes.com/sites/louiscolumbus/2013/05/06/ten-ways-cloud-computing-is-revolutionizing-manufacturing/#1f8d88c852d8

Remember to Subscribe to our Youtube Channel for the Latest Cloud Computing Tech Jobs, News, and Cloud Shows.


David S. Linthicum is a managing director and chief cloud strategy officer. David is internationally recognized as the worlds No.1 cloud computing industry expert, pundit and thought-leader.

(Disclosure: David Linthicum’s views in the blogs, video shows and podcasts are his OWN and are NOT financially sponsored by Nelson Hilliard)

Connect with David on LinkedIn and Twitter:



At Nelson Hilliard we specialise in cloud technologies, sourcing the top 20% of cloud professionals inspired to work for you through our specialised marketing and profiling. If you are interested in having a quick talk to me regarding your employment needs please feel free to reach out.

You can also check my availability and book your 15 minute discovery call here.

Brad Nelson